Table 1.
Power for quantitative incidence and prognosis with non-genetic confounding
| Genetic correlation | 0 | 0 | 0.25 | 0.25 | 0.45 | 0.45 | −0.25 | −0.25 | −0.45 | −0.45 |
|---|---|---|---|---|---|---|---|---|---|---|
| Adjustment | No | Yes | No | Yes | No | Yes | No | Yes | No | Yes |
| All SNPs not affecting prognosis | 5.12 | 5.00 | 5.25 | 5.06 | 5.42 | 5.23 | 5.05 | 5.04 | 5.02 | 5.10 |
| All SNPs affecting incidence but not prognosis | 7.24 | 5.03 | 9.59 | 6.06 | 12.5 | 9.15 | 5.93 | 5.65 | 5.38 | 6.85 |
| SNP with highest type-1 error | 33.0 | 5.7 | 61.7 | 19.1 | 87.9 | 63.7 | 20.0 | 15.0 | 10.8 | 28.5 |
| Family-wise type-1 error | 22.3 | 5.5 | 61.0 | 12.8 | 94.8 | 53.4 | 12.1 | 10.0 | 6.8 | 16.3 |
| All SNPs affecting prognosis | 19.5 | 16.7 | 18.7 | 18.0 | 16.6 | 17.2 | 19.3 | 13.8 | 18.7 | 10.9 |
| All SNPs affecting incidence and prognosis | 20.3 | 16.5 | 16.7 | 16.5 | 10.0 | 11.9 | 21.3 | 13.1 | 20.6 | 8.38 |
| SNP with greatest increase in power | 6.6 | 39.2 | 18.0 | 50.1 | 34.8 | 59.5 | 6.9 | 34.8 | 5.2 | 14.9 |
| SNP with greatest decrease in power | 72.3 | 19.9 | 75.0 | 41.9 | 20.4 | 12.1 | 93.6 | 19.2 | 96.1 | 22.0 |
Estimates shown as % with over 1000 simulations of 100,000 independent SNPs. Five thousand SNPs have effects on incidence only, 5000 on prognosis only and 5000 on both incidence and prognosis. Heritability of both incidence and prognosis is 50% with the genetic correlation shown over all SNPs. Common non-genetic factors explain 40% of variation in both incidence and prognosis. Rows 2–5 show type-1 error rates. All SNPs, mean power across the relevant SNPs. Family-wise error, probability of at least one SNP with effect on incidence but not on prognosis having . SNP with greatest increase (decrease) in power compares the adjusted analysis to the unadjusted